The ability to predict individual tumor radiosensitivity is central to the development of personalized treatment strategies in radiation oncology. In preliminary studies, we developed and biologically validated a systems biology model of tumor radiosensitivity in 48 human cancer cell lines. Furthermore, using the proposed network hubs an accurate gene expression assay to predict cellular radiosensitivity was developed. The assay was then clinically validated in three independent pilot cohorts totaling 118 patients with esophageal, rectal and head and neck cancer treated with concurrent radio-chemotherapy. The purpose of this proposal is to validate, standardize and optimize the gene expression model in a prospective clinical trial. Our clinical hypothesis is that predicted radiosensitivity will be significantly correlated with clinical response to standard preoperative concurrent radiochemotherapy in esophageal and rectal cancer patients. To test this, we propose a prospective trial that will enroll 36 patients with esophageal or rectal cancer that are deemed adequate candidates for preoperative treatment with concurrent radiochemotherapy. All patients will undergo surgical resection 6-8 weeks after completion of preoperative treatment. The main clinical endpoint is pathological response. The translational endpoint is predicted radiosensitivity (SF2). Predicted radiosensitivity will be derived from the 10 gene expression model previously developed and described in the proposal. Gene expression will be derived from microarrays obtained from RNA isolated from biopsies at the time of initial evaluation. We will correlate predicted radiosensitivity (as a continuous variable) to pathological response and determine whether a statistically significant difference in predicted radiosensitivity exists between responders and non-responders. The implications of a successful radiosensitivity predictive assay are broad and significant since it will allow better selection of patients for radiation protocols and thus could potentially improve the ability of physicians to individualize therapy. For example patients with radioresistant tumors may be spared the side effects of radiotherapy if they are unlikely to respond to radiotherapy. In summary, this technology might be central in personalized medicine in radiation oncology. Public Health Relevance: This project will validate a novel technology that has been developed to predict which patients with esophageal or rectal cancer will respond to treatment with radiation and chemotherapy. The relevance of this technology is significant as it may enhance the ability of physicians to personalized treatments for their patients.